PARIS, Sept 2 — Artificial intelligence is now able to generate music in a supposedly “natural” way, based on a theme or a predefined ambiance. While artists can use this kind of AI to make personalised creations, anyone and everyone can use it to accompany podcasts, presentations or animations.
It’s now possible to create musical loops thanks to automatic machine learning. In fact, several online services let people create custom music, whether for personal or commercial projects. This is notably the case of Soundraw and Mubert.
In each case, the creation process is roughly the same. The user first selects an ambiance or an emotion, then specifies the type of rhythm and its speed, before finally choosing the duration of the desired loop. The tool then automatically generates musical patterns thanks to an artificial intelligence system that has “learned” to adapt, through repeated use, to suit each requested mood. Each user is then free to rework these tracks. These online solutions generally offer services suitable for amateurs and professionals, with subscription packages depending on the number of tracks created.
Even more impressively, in 2020, it was demonstrated that an artificial intelligence programme could also create an entire song — lyrics and music — with little or no human intervention. This technology, called Jukebox, is the work of OpenAI, a company specialising in artificial intelligence and whose co-founder is none other than Elon Musk, who also helms Tesla and SpaceX. — ETX Studio
